Mistral AI has taken another large step from European challenger to member of the enterpriseAI establishment, with its models to be integrated into Cloudera’s hybrid data platform.
The deal gives Cloudera customers the ability to use Mistral models with their own governed data across public clouds, private systems and fully disconnected networks. Cloudera says its platform manages 30 exabytes of data, equivalent to 30-billion gigabytes, much of it belonging to large organisations with strict security or regulatory requirements.
It is the latest in a succession of deals that has placed Mistral on many of the platforms through which companies buy and deploy artificial intelligence. Microsoft, Amazon, Google, Snowflake and Databricks began bringing its models onto their platforms during 2024. IBM added Mistral Large 3 to watsonx in December 2025. Mistral models are now available through Microsoft Azure AI, Amazon Bedrock, Google Cloud Vertex AI and IBM watsonx.
Cloudera brings another type of customer into reach: organisations whose data extends across cloud services and their own data centres, including systems kept entirely offline. These air-gapped environments are used where an outside network connection would create an unacceptable security risk.
“EnterpriseAI is entering a new phase where organisations need more than access to powerful models, they need the freedom to unlock specialised intelligence using their data, on their terms,” said Abhas Ricky, chief business officer and general manager for applied AI at Cloudera. “Together with Mistral, we are giving enterprises the ability to run AI where their data lives, customise it with their own intellectual property, and maintain control over their data, infrastructure, and economics. That combination of intelligence and control is essential to moving AI from experimentation into production.”
The partnership comes two days after Mistral announced that it had raised €3-billion at a valuation above €21-billion. That financial backing and the growing list of distribution partners give it a stronger chance of competing for corporate spending dominated by US developers such as OpenAI, Anthropic and Google.
Mistral was founded in Paris in 2023 by former researchers from Meta and Google DeepMind. It built its reputation on efficient models, open-weight releases and a distinctly European argument for technological independence from the largest American AI companies.
Open-weight models allow customers to download and adapt the underlying model parameters. Companies can run them on infrastructure they control, subject to the relevant licence, instead of sending every prompt and document to an outside chatbot service.
That difference has become commercially useful as governments and regulated industries ask where their information is processed, who can access it and which country’s laws apply. “Sovereign AI”, one of the industry’s favourite new labels, describes the ability of a country or organisation to keep control of the information and technology used for its AI systems.
The phrase often carries a thick coating of marketing. In the Cloudera deal, it has a practical meaning. A bank could run a Mistral model next to records held in its own data centre. A government agency could apply it to protected documents inside a disconnected network. A manufacturer could place a smaller model at a plant where a slow or unreliable connection makes a remote cloud service impractical.
Customers will gain access to Mistral models for reasoning, chat, coding, document processing and voice. They will also be able to use Mistral Forge, introduced in March, to build models grounded in their own institutional knowledge.
Forge is aimed at organisations that want deeper customisation than attaching a chatbot to a folder of documents. It allows a company to train or adapt a model around its specialist knowledge and working processes while keeping the data in a controlled environment.
For Cloudera, Mistral adds another model provider to an enterpriseAI ecosystem designed to prevent customers being tied to one supplier. For Mistral, Cloudera supplies access to the hardest end of the corporate market, where sensitive records are spread across several generations of infrastructure.
That part of the market rarely attracts the excitement generated by a new consumer chatbot. It is also where a great deal of enterprise technology spending takes place. Large companies tend to keep systems for years, and many have accumulated data across acquisitions, regulatory archives and applications that cannot simply be moved into one public cloud.
Running AI close to the data can reduce the cost and risk of transferring huge collections elsewhere. It can also help a company keep one set of access rules and audit controls around the information. The customer still has to test the model and manage who may use it, while checking the output. Sovereignty does not cure hallucinations or weak implementation.
The economics will be another attraction. Companies relying entirely on a public application programming interface pay whenever their software calls the model. Heavy use can turn a promising AI experiment into a large and unpredictable bill. Cloudera customers will be able to choose the infrastructure and deployment method suited to each workload, including running inference on their own equipment.
Inference is the stage at which a trained model processes a request and produces an answer. Placing it within a company’s environment gives the organisation greater control over performance and cost, provided it has the computing resources and skills to operate the system.
The companies also plan to work on AI at the edge, where models run close to the equipment or location generating the data. Possible settings include factories and remote facilities with limited connectivity. The announcement gives no timetable for those capabilities, and says further integrations will be introduced over time.
Mistral’s presence across the leading cloud and data platforms gives corporate buyers a European alternative without asking them to abandon the technology estates they already use. Cloudera extends that reach to data that companies have kept behind their own walls.
“Our mission is to make frontier AI available to enterprises without requiring them to give up control over their data, infrastructure, or intellectual property,” said Kamal Brar, senior vice-president of partnerships and alliances at Mistral. “Cloudera manages some of the world’s most valuable enterprise data estates, making this partnership a powerful opportunity to bring our technology directly to where that data lives. Together, we can give customers the ability to build AI that reflects their own data and expertise and deploy it securely wherever their business requires.”